Truck–Drone Collaborative Systems for Scalable Last-Mile Parcel Delivery

Abstract

Truck–drone collaborative delivery (TDCD) has emerged as a promising approach to improve the efficiency and sustainability of last-mile logistics by combining the complementary capabilities of ground vehicles and unmanned aerial systems. Despite growing research interest, many existing studies rely on simplified operational assumptions, make limited use of empirically derived drone performance data, and give comparatively less attention to deployment conditions and user acceptance. This dissertation addresses these limitations through an integrated investigation of drone performance modeling, coordinated optimization, and behavioral adoption. This thesis develops a comprehensive, data-driven framework for the design and optimization of TDCD systems. At the operational level, an optimization framework is proposed to jointly optimize truck stop selection, customer assignment, and drone fleet allocation under realistic constraints, including limited fleet size and energy capacity. To capture drone performance more accurately, regression-based models are developed to estimate delivery time and energy consumption, and are integrated into the optimization framework. The framework is evaluated using real-world route data derived from the Amazon Last Mile Routing Research Challenge dataset (2021), ensuring practical relevance. Computational experiments and sensitivity analyses demonstrate that the proposed TDCD framework can significantly outperform conventional truck-only delivery systems, achieving substantial reductions in delivery completion time and total energy consumption while revealing diminishing returns beyond certain fleet sizes and infrastructure configurations. The results highlight the importance of coordinated decision-making across routing, resource allocation, and infrastructure design to maximize system performance and scalability. Beyond operational optimization, this thesis examines the behavioral dimension of drone delivery adoption through a questionnaire-based study grounded in an extended UTAUT2 framework. The findings identify performance expectancy, social influence, and perceived risk as key determinants of Canadians’ adoption intentions. By integrating optimization modeling, performance evaluation, and user acceptance analysis, this research advances understanding of TDCD and provides practical insights for developing scalable and sustainable last-mile delivery systems.